kafka-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Swapnil Ghike (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (KAFKA-495) Handle topic names with "/" on Kafka server
Date Wed, 05 Sep 2012 01:18:07 GMT

    [ https://issues.apache.org/jira/browse/KAFKA-495?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13448374#comment-13448374
] 

Swapnil Ghike commented on KAFKA-495:
-------------------------------------

So I am planning to insert the following regex check.

Unix constaints - 
1. Ban / (both not allowed in unix, / is used as a path name component separator and \0 terminates
file names in unix.)

Zookeeper constaints:- 
1. The null character (\u0000) cannot be part of a path name. (This causes problems with the
C binding.)
2. The following characters can't be used because they don't display well, or render in confusing
ways: \u0001 - \u0019 and \u007F - \u009F.
3. The following characters are not allowed: \ud800 -uF8FFF, \uFFF0-uFFFF, \uXFFFE - \uXFFFF
(where X is a digit 1 - E), \uF0000 - \uFFFFF.
4. . and .. Are not allowed as filenames. Zookeeper supports only absolute filepaths.
5. The token "zookeeper" is reserved.

In addition, I suggest the following two - 
1. Ban leading and trailing whitespaces (because they confuse GUI users) - also takes care
forbidding filenames composed of only whitespaces.
2. Ban leading . And .. Because unix hides those files.

There are other cases where unfortunate errors may happen in ways not related to kafka or
zookeeper, and programmers would need to take care about them.
                
> Handle topic names with "/" on Kafka server
> -------------------------------------------
>
>                 Key: KAFKA-495
>                 URL: https://issues.apache.org/jira/browse/KAFKA-495
>             Project: Kafka
>          Issue Type: Bug
>    Affects Versions: 0.7, 0.8
>            Reporter: Neha Narkhede
>            Assignee: Swapnil Ghike
>              Labels: bugs
>             Fix For: 0.8, 0.7.1
>
>         Attachments: kafka-495-v1.patch
>
>
> If a producer publishes data to topic "foo/foo", the Kafka server ends up creating an
invalid directory structure on the server. This corrupts the zookeeper data structure for
the topic - /brokers/topics/foo/foo. This leads to rebalancing failures on the consumer as
well as errors on the zookeeper based producer. 
> We need to harden the invalid topic handling on the Kafka server side to avoid this.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message